With an increasing number of cloud applications, the scale of cloud deployment is getting larger, data grow in a magnitude of ten millions, and a variety of cloud storage products come into being. However, these cloud storage products should have one thing in common that expired data are constantly removed to make space for storing new data, thus constantly recycling storage space to save costs. In this case, in face of user requirements that some important data are not desired to be removed even they are expired and are required to be reserved for a longer time, we need a new solution to solve this problem.
Now for some cloud storage products, users can mark corresponding data with labels as required, so as to directly query required data based on label information later, and if important data are required to be preserved for a long time, expired data need to be manually removed by human to make space for storing new data; for some cloud storage products, any type of data can be stored which is too general, so that video data storage performances of large data stream are affected, and expired data also need to be manually removed to make space. The above two cases both need unwanted data to be manually removed by human, whereas manual operations often cause some problems that important data are mistakenly removed and cannot be restored or expired data are not removed which leads to resource waste, thus operation and maintenance staff are required to be very careful and a maintenance process will naturally slow down, thereby increasing maintenance costs.